All the Best Polls Agree with Me: Bias in Evaluations of Political Polling

Total Recall? Examining the Accuracy of Poll Recall during an Election Campaign

“…Hence, even in situations where people encounter the same polls and recall results in an unbiased manner, they might still indicate different numbers, because they weight polls differently depending on whether they help them achieve their "directional" motivations. Recent studies have shown that people view polls as less credible when the poll result contradicted their own view, (for example, Kuru et al, 2017;Madson & Hillygus, 2020) while the ideological alignment of the news source does not seem to influence perceived poll credibility (Kuru et al, 2017). Disentangling the potential sources and mechanisms of poll selection and processing is demanding, perhaps evidenced by the very negligible effects of partisan sentiment found in our analyses, but it is nevertheless a promising task for future research.…”

“…other factors, like the general tone of political media coverage, political preferences, and perceived opinions in people's personal social environments are controlled for (Zerback et al, 2015). Despite this considerable interest in poll effects, our knowledge of how they come about is still quite limited, although more recent work has also focused on the perception and assessment of polls (e.g., Madson & Hillygus, 2020). For example, Mutz (1997) asserts a need for understanding the social psychological processes behind poll effects since the majority of empirical studies only focus on behavioral outcomes of poll exposure, particularly participation in elections and voting decisions.…”

Meta-Cognitive Reasoning is Associated with Distrust in Polls, Independent of Motivated Reasoning

“…To our knowledge Study 1 is the first to establish a robust relationship between perceived respondent honesty and trust in the poll. Our work builds upon the nascent literature on public distrust in polls (Kuru et al, 2017;Madson & Hillygus, 2019) by not only demonstrating the associations between poll distrust and motivated reasoning, meta-cognitive reasoning, and ideology, but also by explicating the relative strength of those associations. Study 1 suggests that the hitherto unstudied attribution of respondent dishonesty, and previously discounted effect of perceptions of poll source credibility (e.g., Kuru et al, 2017), in fact play a larger role in poll distrust than political ideology and motivated reasoning.…”

“…We divide explanations of mistrust in polls into two broad categories: motivated reasoning and meta-cognitive reasoning. Motivated reasoning, in the context of polls, reflects the relationship between trust in a poll and agreement with or endorsement of the poll's results (Kuru et al, 2017;Madson & Hillygus, 2019). Relatedly, we also examine the potential main effect of ideology on trust in polls, as American conservatives display lower levels of trust in the media in general (Middleton, 2018;Mourão et al, 2018).…”

“…One potential psychological process is motivated reasoning, the well-documented phenomenon of individuals rejecting information that conflicts with their prior beliefs and strongly held attitudes (Brashier & Marsh, 2020;Epley & Gilovich, 2016;Jerit & Barabas, 2012). Individuals reject poll results that go against their personally held opinions and desired political outcomes (Tsfati, 2001), and political attitudes become more polarized when individuals view polls that align with their prior attitudes (Madson & Hillygus, 2019). The effect of motivated reasoning on perceived poll credibility is also stronger among individuals with high levels of political knowledge (Kuru et al, 2017), further suggesting a relationship between ideological motives and reactions to poll results.…”